Phylogenetically and Spatially Close Marine Harbour Divergent Bacterial Communities

Cristiane C. P. Hardoim1, Ana I. S. Esteves1, Francisco R. Pires2, Jorge M. S. Gonc¸alves3, Cymon J. Cox4, Joana R. Xavier2,5, Rodrigo Costa1* 1 Microbial Ecology and Evolution Research Group, Centre of Marine Sciences, University of Algarve, Faro, Algarve, Portugal, 2 Centro de Investigac¸a˜o em Biodiversidade e Recursos Gene´ticos, Laborato´rio Associado, Po´lo dos Ac¸ores, Departamento de Biologia da Universidade dos Ac¸ores, Ponta Delgada, Ac¸ores, Portugal, 3 Fisheries, Biodiversity and Conservation Research Group, Centre of Marine Sciences, University of Algarve, Faro, Algarve, Portugal, 4 Plant Systematics and Bioinformatics, Centre of Marine Sciences, University of Algarve, Faro, Algarve, Portugal, 5 Centre for Advanced Studies of Blanes, Girona, Spain

Abstract Recent studies have unravelled the diversity of -associated bacteria that may play essential roles in sponge health and metabolism. Nevertheless, our understanding of this microbiota remains limited to a few host species found in restricted geographical localities, and the extent to which the sponge host determines the composition of its own microbiome remains a matter of debate. We address bacterial abundance and diversity of two temperate marine sponges belonging to the Irciniidae family - Sarcotragus spinosulus and variabilis – in the Northeast Atlantic. Epifluorescence microscopy revealed that S. spinosulus hosted significantly more prokaryotic cells than I. variabilis and that prokaryotic abundance in both species was about 4 orders of magnitude higher than in seawater. Polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE) profiles of S. spinosulus and I. variabilis differed markedly from each other – with higher number of ribotypes observed in S. spinosulus – and from those of seawater. Four PCR-DGGE bands, two specific to S. spinosulus, one specific to I. variabilis, and one present in both sponge species, affiliated with an uncultured sponge-specific phylogenetic cluster in the order Acidimicrobiales (Actinobacteria). Two PCR-DGGE bands present exclusively in S. spinosulus fingerprints affiliated with one sponge-specific phylogenetic cluster in the phylum Chloroflexi and with sponge-derived sequences in the order Chromatiales (Gammaproteobacteria), respectively. One Alphaproteobacteria band specific to S. spinosulus was placed in an uncultured sponge-specific phylogenetic cluster with a close relationship to the genus Rhodovulum. Our results confirm the hypothesized host-specific composition of bacterial communities between phylogenetically and spatially close sponge species in the Irciniidae family, with S. spinosulus displaying higher bacterial community diversity and distinctiveness than I. variabilis. These findings suggest a pivotal host-driven effect on the shape of the marine sponge microbiome, bearing implications to our current understanding of the distribution of microbial genetic resources in the marine realm.

Citation: Hardoim CCP, Esteves AIS, Pires FR, Gonc¸alves JMS, Cox CJ, et al. (2012) Phylogenetically and Spatially Close Marine Sponges Harbour Divergent Bacterial Communities. PLoS ONE 7(12): e53029. doi:10.1371/journal.pone.0053029 Editor: Tilmann Harder, University of New South Wales, Australia Received August 15, 2012; Accepted November 19, 2012; Published December 27, 2012 Copyright: ß 2012 Hardoim et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was financed by the Portuguese Foundation for Science and Technology (FCT - http://www.fct.pt) through the research project PTDC/MAR/ 101431/2008. CCPH has a PhD fellowship granted by FCT (Grant No. SFRH/BD/60873/2009). JRX’s research is funded by a FCT postdoctoral fellowship (grant no. SFRH/BPD/62946/2009). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected]

Introduction knowledge of microbial diversity in marine sponges, with more than 25 bacterial phyla detected in sponges by this means [8,9]. Marine sponges have been the focus of increasing microbiology Taken together, these features foreshadow marine sponge research interest mainly because of their symbiotic association with holobiomes as valuable reservoirs of microbial genetic and abundant and diverse bacteria and production of biologically metabolic diversity of potential use in biotechnology. active secondary metabolites [1,2]. For so-called High Microbial Despite such remarkable advances, current understanding of Abundance (HMA) sponges, it has been shown that up to 38% of symbiont community structure in marine sponges remains re- sponge wet weight is composed of bacterial cells [3], and that such stricted to certain regions and host species [1]. This holds true for bacterial abundance surpasses that of seawater by 2 to 4 orders of species within the family Irciniidae (Demospongiae, Dictyocer- magnitude [1,4–6]. It has been suggested that HMA sponges atida), from which the majority of surveys undertaken so far have harbour several bacteria involved in the production of secondary been limited to tropical latitudes and to the species Ircinia felix, metabolites, which might, for example, improve protection against I. strobilina, and I. ramosa [8,10–17]. Electron microscopy analyses predation of the sponge host [1]. The synthesis of bioactive unveiled abundant and diverse bacterial morphotypes in I. felix compounds derived from sponge-microbe associations has already [10–12], whereas five [15] and seven [17] bacterial phyla were been reported for 26 of the 92 families in the Demospongia [7], revealed in association with I. strobilina by cloning-and-sequencing the most diversified class of the phylum Porifera. Currently, the of 16S rRNAgene fragments. By means of high-throughput use of high-throughput sequencing technology is extending our sequencing, sixteen phyla and 1199 bacterial operational taxo-

PLOS ONE | www.plosone.org 1 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges nomic units (OTUs) at 95% sequence similarity were found in sequences of the subunit I of the mitochondrial cytochrome C association with I. ramosa [8], highlighting the complexity of the oxidase (CO1) gene obtained for all specimens (accession numbers Ircinia-associated ‘‘bacteriome’’. The detection of Acidobacteria, HE797930 to HE797937) showed no intraspecific variation Alpha- and Gammaproteobacteria in adult, larva and juvenile samples among our sequences of I. variabilis or S. spinosulus, whereas of I. felix [11] supports the hypothesis that a portion of this a 4.7% genetic distance (p-distance) was found between our microbiota might be vertically transmitted throughout successive sequences of these two species. Genetic distances between our S. host generations. spinosulus sequences and those available on NCBI GenBank ranged Conversely, the microbial ecology of temperate irciniids remains from 0 to 0.6%, whereas for I. variabilis a distance of 0.5% was underexplored. Only recently a study first approached the observed between our sequences and those of I. variabilis/fasciculata diversity of bacterial communities in Mediterranean Ircinia spp. - collected in the Northwestern Mediterranean. Phylogenetic namely I. variabilis, I. fasciculata and I. oros - revealing eight bacterial reconstructions based both on Maximum Likelihood and Bayesian phyla across these hosts and species-specific OTUs [18]. Because inference confirmed the identification of our sponge specimens. of their global distribution, encompassing both tropical and Indeed, I. variabilis and S. spinosulus CO1 sequences sampled in this temperate species, Irciniidae sponges constitute a valuable taxon study formed well-supported clades with CO1 sequences from for the study of the ecology and evolution of symbiotic relation- Ircinia spp. and Sarcotragus spp. retrieved elsewhere (Fig. 1). ships. In addition, a great variety of cytotoxic compounds has been retrieved from Irciniidae species, which indicates they are poten- Counting of Heterotrophic Culturable Bacteria tially of high biotechnological importance. [19–24]. Furthermore, The colony forming units (CFU) counts of heterotrophic two studies performed with the temperate I. muscarum and I. bacteria on marine agar revealed no significant difference variabilis described the production of cyclic peptides by cultivated (p.0.05) between sponges species, with 3.2162.036106 CFU bacteria [25,26], whereas psymberin – which resembled the and 1.6360.616106 CFU g21 of fresh sponge for S. spinosulus and pederin family of polyketides – was recovered from Psammocinia sp. I. variabilis, respectively. and shown to have a bacterial symbiont origin [19,27]. In this context, addressing microbial diversity and distribution in Epifluorescence Microscopy widespread and chemically complex marine sponges is not only Analyses showed that S. spinosulus harboured significantly relevant to the study of symbiosis and co-evolutionary relation- (p,0.05) higher number of prokaryotic cells (average of ships, but also bears importance to our understanding of the extent 1.3761010 cells g21 of fresh sponge), as surveyed in this study, of marine genetic and metabolic resources. when compared to I. variabilis (average of 3.816109 cells g21 of In light of the recent evidence for divergent bacterial fresh sponge). The abundance of prokaryotic cells in surrounding communities across different sponge species or even specimens seawater (average of 4.636105 cells mL21) was significantly [9,28], a feature that has also been observed for other eukaryotes (p,0.05) lower than in both sponge species (Fig. 2). that support complex bacterial consortia [29–31], this study uses a stringent experimental design to test the hypothesis of host- PCR-DGGE Analysis of Bacterial Communities specific assemblages of dominant symbionts in marine sponges. To (i) Bacteria PCR-DGGE profiles. The bacterial PCR- this end, we address bacterial abundance and diversity in the DGGE profiles of S. spinosulus were characterized by 5 dominant temperate marine sponges Sarcotragus spinosulus Schmidt, 1862 and bands and a large number of fainter bands (16 to 30) whereas Ircinia variabilis Schmidt, 1862 (Demospongiae, , those of I. variabilis comprised 1 dominant band in addition to 5 to Irciniidae), two closely related species that co-exist in spatial 26 fainter bands (Fig. 3a, Table S1). Seawater DGGE profiles proximity at the coast of the Algarve (southern Portugal), a region showed 7 dominant bands and a large number of fainter bands (28 with a Mediterranean climate located in the Northeast Atlantic. to 30). While the similarity within seawater and S. spinosulus We use the 16S rRNA gene as a phylogenetic marker in replicates was high, profiles of I. variabilis specimens displayed large polymerase chain reaction – denaturing gradient gel electropho- within-replicate heterogeneity (Fig. 3a). Clearly contrasting profiles resis (PCR-DGGE) analyses of the domain Bacteria, the phylum were observed between seawater and sponge samples, and Actinobacteria and the class Alphaproteobacteria in these hosts, thus between both sponge species. Indeed, the UPGMA cluster analysis allowing the inspection of bacterial community structures across (Fig. S2a) revealed two main groups, one formed exclusively by all different taxonomic ranks with concomitant focus on abundant sponge specimens and other containing only seawater samples. and biotechnologically relevant sponge-associated microorganisms These two groups shared less than 10% similarity whereas S. [7,32,33]. Phylogenetic analysis of dominant bacterial populations spinosulus and I. variabilis PCR-DGGE profiles shared around 20% (i.e. PCR-DGGE bands) consistently and specifically found in similarity. Ordination via canonical correspondence analysis association with these species is performed, and their status as (CCA) of the DGGE band data and environment variables ‘‘sponge-specific bacterial clusters’’ [34] is verified. We also revealed that sponge species and seawater significantly influenced determine the degree of dissimilarity between sponge-associated band variation in the DGGE profiles (p,0.05, Fig. 3b). The bacterial communities and that of their neighbouring bacterio- horizontal axis of the diagram, which accounted for 54.8% of the plankton. To assure accurate identification of the target sponges, dependent (i.e. DGGE bands) – independent (i.e. sample classes) we infer host phylogenies based on cytochrome oxidase gene variables correlations, mainly distinguished the sponges S. sequence relationships. This is the first study of bacterial spinosulus and I. variabilis from seawater (Fig. 3b). The vertical axis community structure and diversity in North Atlantic Irciniidae. grouped replicates from sponge S. spinosulus clearly apart from those observed in I. variabilis. Results (ii) Actinobacteria PCR-DGGE profiles. The actinobacter- ial PCR-DGGE profiles of S. spinosulus consisted of few (2 to 4) Sponge Identification strong bands along with more than 4 detectable bands, while those Sponge specimens (Fig. S1) were identified as Sarcotragus of I. variabilis comprised 1 to 3 dominant bands along with 1 to 5 spinosulus and Ircinia variabilis based on macro- and microscopic fainter bands (Fig. 3c). Significant reduction in actinobacterial analyses using morphological criteria. Analysis of 636 bp-long diversity and richness were determined for the latter species in

PLOS ONE | www.plosone.org 2 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 1. Phylogenetic inference of the Irciniidae family based on the cytochrome oxidase gene, subunit 1. The Maximum Likelihood tree (-ln likelihood: 1383.921591) is shown, with sequences retrieved in this study highlighted in bold. Numbers at tree nodes are bootstrap values and posterior probabilities calculated in Maximum Likelihood and MCMC Bayesian analyses, respectively, and values above 70/0.95 are shown. doi:10.1371/journal.pone.0053029.g001 comparison with the former (Table S1). A large heterogeneity was spinosulus which clustered with I. variabilis specimens (Fig. 3f, S1c). observed within I. variabilis profiles. In comparison with the sponge Nevertheless, CCA showed that all factors significantly (p,0.05) fingerprints, the seawater PCR-DGGE profiles displayed higher influenced the patterns of band distribution in alphaproteobacter- diversity of bands, especially against I. variabilis profiles (Table S1), ial PCR-DGGE profiles. Ordination via CCA revealed that 63% and contained 3 strong bands along with more than 6 fainter of total PCR-DGGE band – independent variables correlations bands, and much lower within-replicate variability (Fig. 3c). Two was explained by the horizontal axis of the diagram, which mainly main groups were revealed by cluster analysis, one formed differentiated S. spinosulus from seawater (Fig. 3f), whereas the exclusively by all sponge specimens and other containing only residual variability in the vertical axis of the diagram (37%) seawater samples (Fig. S2b). These two groups shared about 10% discriminated most I. variabilis from seawater and S. spinosulus similarity. However, there was also a clear difference between the samples (Fig. 3f). profiles of both sponge species, which shared c. 20% similarity according to cluster analysis. Ordination via CCA discriminated Analysis of Sequences of Dominant and Discriminating both sponge species and seawater across the horizontal axis of the PCR-DGGE Bands diagram, which represented around 60% of the overall PCR- (i) Bacteria PCR-DGGE bands. Three dominant bands DGGE – sample correlations (Fig. 3d). All independent variables labelled 1, 2 and 4 (see arrows in Fig. 3a) were exclusively found in (i.e. the sample classes seawater, S. spinosulus and I. variabilis) all replicates of S. spinosulus. Bands 1 and 4 were directly sequenced significantly (p,0.05) affected the PCR-DGGE banding patterns whereas band 2 was subjected to cloning and sequencing. From (Fig. 3d). band 1, one sequence was retrieved and affiliated with the (iii) Alphaproteobacteria PCR-DGGE profiles. The Al- Actinobacteria order Acidimicrobiales. The phylogenetic analysis phaproteobacteria profiles of S. spinosulus contained 1 to 3 dominant showed that this band affiliated with an uncultured and apparently bands, in addition to more than 7 detectable fainter bands, diverse lineage containing sponge-derived bacterial sequences of whereas the profiles of I. variabilis revealed 1 to 3 strong and fainter worldwide origin (Fig. 4). Further, two clones were sequenced bands (Fig. 3e). Significantly greater richness, diversity and from band 2 and found to be highly alike, with 5 different evenness were found for S. spinosulus alphaproteobacterial PCR- nucleotides between them. They were assigned to the Gammapro- DGGE profiles in comparison with those of I. variabilis (Table S1). teobacteria order Chromatiales. These sequences belonged to a well The seawater profiles showed 2 strong bands along with more than supported Chromatiales phylogenetic clade containing uncultured 6 fainter bands. The variability within sponge specimens and bacteria retrieved exclusively from marine sponges sampled in among sponge species was relatively high. Conversely, seawater several geographical localities (Fig. 5). From band 4, one sequence samples displayed highly homogeneous profiles (Fig. 3e). Cluster was obtained and classified in the Chloroflexi phylum. Phylogenetic analysis revealed a clear separation between seawater and sponge analysis revealed that this sequence belonged to a sponge-specific samples and high similarity scores for the former (Fig. S2c). The bacterial phylogenetic cluster as determined by Simister et al. [34] latter grouped into two further clusters in which the visible, higher (Fig. 6). A dominant band labelled 3 (Fig. 3a) was found in all degrees of within-sample variability could be numerically depicted sponge specimens. Two identical sequences were recovered and (Fig. S2c). Sample outliers were detected, as one replicate from I. assigned to the Actinobacteria order Acidimicrobiales. They also variabilis grouped with a cluster dominated by three S. spinosulus affiliated with an uncultured sponge-specific lineage previously samples, and the same effect was observed for one replicate from S. suggested by Simister et al. [34] (Fig. 4). A dominant band labelled

PLOS ONE | www.plosone.org 3 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 2. Epifluorescence counts. Microscopy pictures taken from S. spinosulus (A), I. variabilis (B) and Seawater (C) are shown. Arrows exemplify counted bacterial cells. Values in panel (D) are expressed as means 6 standard errors of log-transformed total cell numbers (TCN). doi:10.1371/journal.pone.0053029.g002

5 (see arrow in Fig. 3a) was exclusively found in all seawater that these sequences affiliated with bacterial phylotypes retrieved samples. This band possessed high discriminating power, as almost solely from marine sponges distributed worldwide. No obviated by its centroid position in the CCA diagram. Two cultured representative isolated from marine sponges has been identical sequences were obtained for this band and assigned to the observed in this cluster (Fig. 7). One band found almost in all Alphaproteobacteria family Rhodobacteraceae (Table 1). They belonged samples in addition to a dominant band found exclusively in to a supported, uncultured bacterial phylogenetic clade containing seawater samples labelled, respectively, 9 and 10 (Fig. 3e) were sequences retrieved solely from seawater (data not shown). subjected to sequencing. One and three sequences were obtained (ii) Actinobacteria PCR-DGGE bands. The dominant from bands 9 and 10, respectively. They all shared high similarity bands labelled 6 and 7 (Fig. 3C) were recovered from three at the primary sequence level (up to 3 nucleotide differences specimens of S. spinosulus and from all I. variabilis specimens, detected), belonged to a phylogenetic cluster containing several respectively. These bands were subjected to cloning and sequenc- uncultured bacterial phylotypes retrieved only from seawater, and ing. Three clones were obtained from each band, which contained affiliated with the family Rhodobacteraceae (Table 1). 2 and 3 dissimilar nucleotides for bands 6 and 7, respectively. All sequences were assigned to the order Acidimicrobiales. Phylogenetic Discussion analysis revealed that these sequences fell into a sponge-specific bacterial phylogenetic clade [34] from which no cultured This survey addresses bacterial abundance, diversity and representative has so far been registered (Fig. 4). specificity in the Atlanto-Mediterranean sponges Sarcotragus (iii) Alphaproteobacteria PCR-DGGE bands. The domi- spinosulus and Ircinia variabilis (Demospongiae, Dictyoceratida, nant band labelled 8 (Fig. 3e) appeared in all specimens of S. Irciniidae). These species are widely distributed along the southern spinosulus. Three identical sequences were recovered and assigned Portuguese coast (http://www.marinespecies.org/porifera/). Both to the order Rhodobacterales. Phylogenetic reconstruction revealed species were initially identified by traditional taxonomic methods.

PLOS ONE | www.plosone.org 4 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

PLOS ONE | www.plosone.org 5 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 3. PCR-DGGE fingerprints and canonical analyses. PCR-DGGE 16S rRNA gene fingerprints of S. spinosulus, I. variabilis and seawater DNA samples generated with ‘‘total-community’’ bacterial primers (A) and specific primer systems for Actinobacteria (C) and Alphaproteobacteria (E). The arrows indicate bands that were excised from DG-gels and sequenced. Corresponding ordination biplots of PCR-DGGE fingerprints and qualitative environmental variables are shown in panels B, D, F. Symbols: m S. spinosulus, & I. variabilis and N Seawater. Labels displayed on the diagram axes refer to the percentage variations of PCR-DGGE ribotypes - environment correlation accounted for the respective axis. The ‘‘star’’ symbols represent the centroid positions of the environmental variables in the diagram. Variables that significantly (p,0.05) influence the bacterial community composition are indicated by an asterisk. doi:10.1371/journal.pone.0053029.g003

However, species within the Order Dictyoceratida to which the such as the members of the Irciniidae family and its relevance in family Irciniidae belongs are, along with the Order Dendroceratida, such studies is likely to rise with the analysis of multiple known as ‘keratose’ sponges, which usually lack a suite of phylogenetic markers in concatenation. morphological features making their classification problematic In the present survey, the abundance of culturable bacteria [35,36]. In recent years, molecular characterization of sponges by associated with S. spinosulus and I. variabilis was similar. It is well- sequencing of standard genetic markers – known as DNA known that many aspects affect bacterial cultivation and the use of barcoding – is being used increasingly as a means to facilitate standard culture media has so far allowed the assessment of only identification and to complement the description of new species a minor fraction (e.g. from 0.1 to 1%) of bacteria associated with [37]. Almost invariably, analyses involve the use of the subunit I of marine sponges [44–46]. This might sharply compromise the the cytochrome C oxidase gene (CO1) [37–39]. The genetic comparative assessment of bacterial abundance in sponges when variation (p-distance) found in our host species’ CO1 gene, i.e. no solely using this technique. To circumvent the limitations inherent intraspecific variation and a 4.7% genetic distance between from cultivation, epifluorescence microscopy was applied to I. variabilis and S. spinosulus, are within the range of values observed estimate abundance by enabling the count of all detectable for other Irciniids using the same marker. In a barcoding study of nucleic-acid containing cells present in the samples. Based on the Indo-Pacific Irciniids, Po¨ppe et al. [40] observed no intraspecific cell counts retrieved with this method, about 3 orders of variation within any of the analysed species, low interspecific magnitude higher than the registered CFU counts, S. spinosulus variation between congeners (0.2–2.7% in Ircinia spp. and 0.2– and I. variabilis can be regarded as HMA sponges, supporting 1.7% in Psammocinia spp.), and higher differentiation levels previous observations obtained for tropical Irciniidae species such between members of the two genera (p-distances of 3.1–5.8%) as I. felix and I. strobilina [12,47,48]. [40]. In a second study comparing the bacterial symbionts in three Bacteria, Actinobacteria and Alphaproteobacteria PCR-DGGE finger- species of Ircinia in the Mediterranean Sea, Erwin et al. [18] found printing revealed a clear difference in bacterial diversity and no intraspecific variation within any of the studied species (nor community composition between sponge and seawater samples. between I. variabilis/fasciculata) and a p-distance of 0.6–1.8% This expected trend has been reported in several previous sponge between the different species. Not unexpectedly we found some microbiology surveys [1,49,50]. In agreement with our results, the genetic distance (p-distance 0–0.6%) between the sequences of our bacterial PCR-DGGE profiles from I. felix collected at two sites in Atlantic specimens and those available on GenBank from Key Largo, Florida, revealed distinct band patterns in comparison Mediterranean specimens. This may indicate some level of genetic with seawater samples [12], whereas bacterial PCR-DGGE isolation and differentiation between conspecific populations profiles of wild and captivated I. strobilina specimens were likewise occurring in these areas as previously observed in other sponge distinct when compared with surrounding seawater and water taxa (e.g. Xavier, et al. [41]). Overall, host phylogenetic inference used in sponge aquaculture, respectively [15]. In contrast with can be a suitable and complementary tool in sponge microbiology species of the genus Ircinia, knowledge of bacterial abundance and studies - as shown in early [36,42] and recent [18,43] reports on diversity in Sarcotragus specimens is virtually nonexistent. Here we host-symbiont co-evolutionary relationships. Its use seems espe- showed S. spinosulus-specific profiles that strongly differed from cially well suited to the study of sponge hosts displaying smooth those found in I. variabilis, with several species-specific PCR- gradients of phylogenetic relatedness or unresolved taxonomies DGGE bands detected and further identified (see below). S.

Table 1. Closest 16S rRNA gene relatives of seawater-derived and ‘‘cosmopolitan’’ PCR-DGGE bands.

Band id (Accession number) RDP closest match1 (Accession number) NCBI closest match2 (Percent similarity, accession number)

5a (HE797944) Uncultured Roseobacter sp. (AY627365) Uncultured Rhodobacteraceae bacterium clone GG101008Clone5 (99%, JN591908) 5b (HE797945) Uncultured Roseobacter sp. (AY627365) Uncultured Rhodobacteraceae bacterium clone GG101008Clone5 (99%, JN591908) 9a (HE797955) Uncultured Roseobacter sp. (AY627365) Uncultured Rhodobacteraceae bacterium clone GG101008Clone5 (99%, JN591908) 10a (HE797957) Uncultured Roseobacter sp. (AY627365) Uncultured Rhodobacteraceae bacterium clone GG101008Clone5 (100%, JN591908) 10b (HE797958) Uncultured Roseobacter sp. (AY627365) Uncultured Rhodobacteraceae bacterium clone GG101008Clone5 (100%, JN591908) 10c (HE797956) Uncultured Roseobacter sp. (AY627365) Uncultured Rhodobacteraceae bacterium clone GG101008Clone5 (100%, JN591908)

1Closest 16S rRNA gene relative using the ‘‘sequence match’’ tool of the Ribosomal Database Project (RDP). 2Closest 16S rRNA gene relatives as determined by the blast-n search in the National Center for Biotechnology Information (NCBI) database. doi:10.1371/journal.pone.0053029.t001

PLOS ONE | www.plosone.org 6 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 4. Phylogenetic inference of Actinobacteria 16S rRNA genes. The modified ARB database generated by Simister et al., [34] used long sequences ($1200 bp) to infer the phylogeny and shorter sequences were added using the ARB parsimony interactive tool. Sequences from the sponge-specific cluster 22 (SC22) [34] along with sequences closely related to band 1 and outgroup sequences were selected for further phylogenetic analysis. The Maximum Likelihood tree (-ln likelihood: 4501,317092) is shown, with sequences retrieved in this study highlighted in bold. Numbers at tree nodes are bootstrap values and posterior probabilities calculated in Maximum Likelihood and MCMC Bayesian analyses, respectively, and values above 70/0.95 are shown. doi:10.1371/journal.pone.0053029.g004 spinosulus exhibited greater bacterial community diversity and spinosulus was found. In a recent survey, Erwin et al. [18] could richness, and homogeneity across individual specimens than the detect bacterial OTUs exclusive to the species I. oros, I. fasciculata or latter. Further, evidence for greater prokaryotic abundance in S. I. variabilis in the Mediterranean Sea. Taken together, these studies

PLOS ONE | www.plosone.org 7 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges hint at a fundamental role of the host in shaping the structure and bacteria phyla associated with a wide variety of marine sponges, promoting diversity of symbiont communities within closely with many sponge-specific clusters identified [6,34,61]. So far, only related sponge hosts. Interestingly, functional equivalence and one Chloroflexi species was isolated from the sponge Geodia spp., evolutionary convergence of symbiont communities have been which also clustered with sequences exclusively obtained from suggested as an evolutionary model applicable to the complex marine sponges [62]. In shallow waters, members of Chloroflexi are sponge microbiota, based on the share of core microbial functions able to fix atmospheric carbon through photosynthesis, and thus between six phylogenetically distant sponge species with different these bacteria could provide carbonaceous compounds to the symbiont community structures [51]. In this context, it is tempting sponge host [62]. Recently, a Chloroflexi bacterium was pointed as to speculate that the less studied Sarcotragus also establishes close the likely producer of a novel non-ribosomal peptide synthase [63]. interactions with selected bacterial communities, which regardless Thus, Chloroflexi strains might play important roles in sponge their degree of distinctiveness might have intrinsic functions like nutrition and defence. those observed for Ircinia spp. [11,15,16]. Future studies addressing Using a taxon-specific fingerprinting approach to the Alphapro- microbial functioning in sympatric and phylogenetically close teobacteria, a dominant symbiont exclusive to S. spinosuls was hosts will certainly shed further light on our current understanding uncovered (‘‘band 8’’ in Fig. 3) and found to be closely related to of symbiont evolution within sponges. an uncultured alphaproteobacterium within the family Rhodobac- We successfully identified several sponge-specific bacterial terales [64]. Sequences representing this symbiont formed a concise populations by PCR-DGGE. Four dominant symbionts - two cluster with sequences retrieved from marine sponges in several from S. spinosulus, one from I. variabilis and one found in both geographical backgrounds in addition to cultured representatives sponge species - were affiliated with an uncultured actinobacterial obtained from different environments such as microbial mats, lineage within the order Acidimicrobiales [52]. Three of these seawater, soil from saltpan, water and marine aquaculture pond (‘‘bands’’ 3, 6 and 7 in Figs. 3 and 4) belonged to a cluster of [65,66] (Fig. 7). This clade contained sequences obtained from sponge-specific sequences collected worldwide and called SC22 by adults of Xestospongia muta and Svenzea zeai along with their Simister et al. [34], whereas the fourth grouped into a different and reproductive material [53,54]. This symbiont is closely related to diverse cluster dominated by sponge-derived bacterial sequences Rhodovulum species, in which many type strains have been mostly (Fig. 4). Within cluster SC22, sequences were obtained from adult retrieved from marine habitats. This genus contains species that of Svenzea zeai and Smenospongia aurea along with their reproductive undertake diverse metabolic pathways such as photoautotrophic, material, which suggests that vertical transmission of this particular photoheterotrophic and chemotrophic and occur mainly in phylotype is likely to occur [53,54]. The same observation was marine and hypersaline environments under oxic, microoxic and made for the cluster formed by the fourth symbiont in the anoxic conditions [67]. The metabolic versatility of Rhodovulum Acidimicrobiales group (‘‘band 1’’ in Fig. 4) and related sequences, species indicate that they are able to use the waste generated by from which two sequences from adult S. zeai and one from its sponges. For instance, ammonia, which is a toxic metabolic waste embryo were found [53]. These results indicate an intimate product that could accumulate within the sponge body, especially pattern of relationship between sponge-associated Acidimicrobiales during low pumping activity, might be used as nitrogen source for and their hosts. The order Acidimicrobiales contains mesophilic and Rhodovulum species [1,67,68]. In addition, some strains of moderate thermophilic species and all members are obligatory Rhodovulum could be involved in nitrogen and sulphur cycling, acidophilic found in iron-, sulphur- or mineral-sulfide rich once they are capable to use dinitrogen, sulphur, sulphite, sulphate environments. Species within this order are capable of ferrous and thiosulfate [67]. Vertical transmission has also been iron and sulphur oxidation and ferric iron reduction [55–58]. documented for members of this genus in marine sponges, [54] However, the physiological properties exhibited by cultivated suggesting Rhodovulum as a likely, relevant constituent of the Acidimicrobiales might not necessarily match those of marine sponge sponge-associated microbiome. symbionts, as these usually share lower relatedness to cultured The present study provides first insights into the bacterial species at the 16S rRNA gene level, and therefore further research abundance and diversity in Atlantic S. spinosulus and I. variabilis.In is needed to unveil the ecology and functioning of these symbionts spite of their sympatric occurrence, the inspected species hosted in marine sponges. bacterial communities that differ from each other and from those A prevailing symbiont found exclusively in S. spinosulus was found in seawater. Interestingly, all bands excised from PCR- affiliated with uncultured Gammaproteobacteria within the order DGGE profiles that were exclusive to sponge samples affiliated Chromatiales [59]. These sequences belonged to a cluster of sponge- with previously identified sponge-specific sequence clusters [34] or specific sequences acquired worldwide (Fig. 5). Among them, adult with potentially novel sponge-specific clusters found in the present sequences from Ircinia felix, Smenospongia aurea and Svenzea zeai were survey. Thus, the approach used here enabled not only straightfor- observed along with sequences from reproductive material of ward assessment of overall trends in bacterial community I. felix, S. aurea and Ectyoplasia ferox [11,53,54]. The order structures, but also direct identification of symbionts of putative Chromatiales encompasses members of the purple sulphur bacteria relevance in association with their hosts, given their dominance that are capable of performing anoxygenic photosynthesis using and consistent patterns of occurrence in the analysed specimens, hydrogen sulphide as electron donor [59]. Furthermore, many and their presumed sponge-specific life histories as inferred by 16S Chromatiales species have been shown to perform fixation of rRNA gene phylogenies. Notably, bacterial phylotypes regarded as molecular nitrogen [59,60]. These functions might be highly ‘‘S. spinosulus-specific’’ or ‘‘I. variabilis-specific’’ in this study shared valuable for sponge survival, and the consistency with which high degrees of resemblance with sponge-derived sequences from members of this group are found in marine sponges at a global other biogeographical settings and/or more distantly related scale indeed suggests that Chromatiales species play an important sponge hosts. This picture, in which bacterial signatures not role in their association with such hosts. shared by co-occurring and taxonomically close sponge species are Another phylotype solely recovered from S. spinosulus was found in disparate sponge hosts and localities, most likely derives affiliated with an uncultured, sponge-specific lineage in the from factors of the host and of the environment – including Chloroflexi phylum, named SC46 by Simister et al. [34] (Fig. 6). vertical transmission vs. environmental acquisition of symbionts, The Chloroflexi is regarded as one of the most abundant and diverse specific habitat preferences and life stages of the host - that

PLOS ONE | www.plosone.org 8 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 5. Phylogenetic inference of Gammaproteobacteria 16S rRNA genes. Tree construction procedure was as described for Figure 4, except that sequences closely related to band 2 were selected as well as sequences from SC155 [34], which were used as outgroup. The Maximum Likelihood tree is shown (-ln likelihood: 2696,593494). doi:10.1371/journal.pone.0053029.g005

PLOS ONE | www.plosone.org 9 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 6. Phylogenetic inference of Chloroflexi 16S rRNA genes. Tree construction procedure was as described for Figure 4, except that sequences from SC46 were selected along with sequences from SC47, which were used as outgroup [34]. The Maximum Likelihood tree is shown (-ln likelihood: 3791,095). doi:10.1371/journal.pone.0053029.g006 cooperatively shape the structure of the sponge-associated associated life histories. The high abundance and species-specific microbiome [1,18]. As a result, complex communities of specific character of these assemblages suggest in-faunal microbial composition at the host species or even specimen level [9,28] with communities as overriding drivers of functioning and of genetic concomitant sharing, across sponge species, of generalist sym- and metabolic diversities in coastal ecosystems. bionts displaying broad host range and/or widespread occurrence [43,69] have often been reported for marine sponges. Here, the Materials and Methods distinct communities observed in S. spinosulus and I. variabilis within the same habitat, along with the detection of symbionts showing Sponge and seawater sampling. Four specimens of Sarco- broad host and geographical ranges as inferred by 16S rRNA gene tragus spinosulus and Ircinia variabilis (Demospongiae, Dictyoceratida, phylogenies, hints at a pivotal role for the host in shaping the Irciniidae) were collected by scuba diving at depths around 15 m structure of its own microbiota while revealing versatile and at Gale´ Alta, Armac¸a˜o de Peˆra (37u 049 09.699N and 8u 199 widespread bacterial phylotypes with apparently intimate sponge- 52.199W) in the coast of the Algarve, Portugal, in June 2010.

PLOS ONE | www.plosone.org 10 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Figure 7. Phylogenetic inference of Alphaproteobacteria 16S rRNA genes. Tree construction procedure was as described for Figure 4, except that sequences close related to band 8 were selected along with sequences from SC84 [34], which were used as outgroup. The sequences shown in a box were isolated from different environments. The Maximum Likelihood tree is shown (-ln likelihood: 3253,686594). doi:10.1371/journal.pone.0053029.g007

Measurements of temperature, oxygen and salinity during the containing natural seawater, transferred into cooling boxes, sampling procedure were 14.6uC, 5.95 mg L21 and 35.11 part per brought to the laboratory within few hours and processed upon million (ppm), respectively. In situ pictures of the specimens were arrival. Four samples of seawater (1L each) from the vicinity of the taken to aid laboratory identification (Fig. S1). The individual sponges (i.e. 1 m above the specimens) were also collected as samples were placed, in situ, separately in plastic bags (type ziplocH) above. Prior to sample processing, the sponge specimens were

PLOS ONE | www.plosone.org 11 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges rinsed with sterile Artificial Seawater (ASW) [70] to remove loosely debris. Aliquots (100 mL) of the resulting supernatants were associated organisms. Voucher samples were preserved in 90% individually fixed in 2.5% glutaraldehyde and the volume was ethanol for taxonomic identification and deposited in the Biology completed to 10 mL with sterile ASW. Seawater samples (9.2 mL) Departments zoological collection of the University of the Azores were also fixed in 2.5% glutaraldehyde. From the fixed material, (DBUA.Por). Because sampling did not involve endangered or 100 mL and 10 mL from sponge and seawater samples, re- protected species and did not occur within privately owned or spectively, were filtered through 0.2-mm-pore-size isoporeTM black protected areas, no specific permits were required for the membrane filters (Millipore, Bellerica, MA, USA). The filter was described field studies. stained with the DNA-binding fluorochrome acridine orange, Sponge identification. Specimens were identified from the mounted on glass slides and analysed with an inverted research analysis of general external morphological characters and internal system microscope IX81 (Olympus Europa GmbH, Hamburg, skeletal features, i.e. thickness, degree of fasciculation and presence Germany) where 25 photos per specimen were taken at random. of foreign debris within the spongin fibres and structure of the Cells with a well-defined edge, usually ranging from 0.2 to 1 mmin collagenous filaments. Genera within the family Irciniidae are diameter when coccoid, or reaching up to 5 mm in length when distinguished by the presence of a cortical armour of sand rod-shaped, were counted and served as proxy for prokaryotic cell (exclusive to Psammocinia), and presence (in Ircinia) versus absence abundance in the samples. Larger objects (.5 mm) that could (in Sarcotragus) of foreign debris within the primary fibres [35]. eventually account for eukaryotic organisms were not considered. Phylogenetic inference of sponge specimens (commonly referred Total prokaryotic numbers were log-transformed and analysed by to as ‘‘sponge DNA barcoding’’) was used to aid species One Way ANOVA using PASW Statistics 18 (SPSS Inc., Chicago, identification by molecular means. PCR amplifications were USA). carried out on sponge total community DNA (see below) targeting Total community DNA extraction. Genomic DNA of the subunit I of the cytochrome oxidase gene with the primers about 0.25 g of internal sponge body was extracted using dgLCO1490 and dgHCO2189 [71]. This fragment (c. 640 bp) UltraCleanH Soil DNA isolation kit (Mo Bio, Carlsbad, CA, encompasses the standard ‘‘barcoding’’ partition [72]. The USA) according to the manufacturer’s protocol. Based on reaction mixture (25 mL) contained 1.5 mL of template DNA preliminary PCR-DGGE assessments, this method led to a more (,20 ng), 1X reaction buffer (Bioline, London, UK), 0.16 mM reproducible depiction of bacterial community structures in the deoxynucleoside triphosphates (dNTPs), 4.0 mM MgCl2, 0.64 mg sponges when compared with a method that employs a cell- mL21 of bovine serum albumin (BSA), 0.24 mM of primers and TM separation treatment prior to DNA extraction (Hardoim et al., 0.625U of BioTaq DNA polymerase (Bioline, London, UK). unpublished results), and was therefore chosen for the purpose of After initial denaturation at 95uC for 3 min., 36 cycles of 45 sec at this study. Seawater samples (500 mL) were filtered through 0.2- 94uC, 60 sec at 51uC and 90 sec at 72uC were carried out. A final mm-pore-size nitrocellulose filters (Millipore, Billerica, MA, USA) extension of 10 min at 72uC was used to finish the reaction. All using a vacuum pump. The filters were cut into small pieces and PCR amplifications were carried out in a MyCycle thermal cycler directly used for DNA extraction as explained above. (Bio-Rad, Hercules, CA, USA). Amplicons were checked after Bacterial PCR for DGGE analysis. A nested PCR-de- electrophoresis on 1% agarose gel under UV light. PCR products naturing gradient gel electrophoresis (PCR-DGGE) approach was with right size were cleaned with Sephadex G50 (GE Healthcare chosen - based on its higher detection sensitivity and reproduc- Bio-Science AB, Uppsala, Sweden) columns, quantified with TM ibility when compared with a one-step amplification protocol in Image Lab Software (Bio-Rad, Hercules, CA, USA), and preliminary assays (data not shown) - to assess the total bacterial subjected to sequencing with the chain termination method in an communities in all samples. Nearly full-length 16S rRNA gene Applied Biosystems 3130 genetic analyser using the forward fragments were amplified with the primer pair F27 and R1492 primer. Closest relatives were searched using the megablast and [75]. The reaction mixture (25 mL) was prepared with 1 mLof blastn algorithms of the National Center for Biotechnology template DNA (,20 ng), 1X Stoffel buffer (Applied Biosystems, Information (NCBI) [73]. Closely related sequences from the 2 Foster, CA), 0.2 mM dNTPs, 3.75 mM MgCl 0.1 mg mL 1 of NCBI and the Sponge Barcoding Project (www.spongebarcoding. 2, BSA, 2% (vol/vol) dimethyl sulfoxide (DMSO), 0.2 mM of each org) databases were used to retrieve representative CO1 sequences primer, and 1.25U of Taq DNA polymerase (Applied Biosystems, for phylogenetic inference (see below). Foster, CA). After initial denaturation at 94uC for 5 min, 30 cycles Plate counting of heterotrophic bacteria. Per sponge of 1 min at 94uC, 1 min at 56uC and 2 min at 72uC were specimen, 2.5 g of fresh internal body was cut and transferred to performed, followed by a final extension for 10 min at 72uC. The a 50 mL screw cap polypropylene tube containing 25 mL of amplicons (1.5 mL) were used as template in a subsequent PCR for Calcium/Magnesium Free Artificial Seawater (CMFASW) [74]. DGGE analysis (20 cycles) using the primer pair F984-GC and The sponge samples were ground with sterile mortar and pestle. R1378 [76]. The PCR mixture and thermal cycling followed the The resulting suspensions were collected and allowed to decant for protocol by Costa et al. [77], with half the quantity of Taq DNA 5 min. Serial 10-fold dilutions were then prepared with sterile ASW and plated in triplicate onto Marine broth (Carl Roth polymerase (1.25 U) per reaction. GmbH+Co, Germany) plus 1.5% agar. The plates were incubated at room temperature (,25uC) and Colony Forming Unit (CFU) PCR of Specific Bacterial Groups for DGGE Analyses counting was performed after 3, 5 and 7 days of incubation. Log- Actinobacteria 16S rRNA gene fragments. The first transformed CFU values fitted the normal distribution and were amplification of the nested PCR was carried out with the primers compared by One Way Analysis of Variance (ANOVA) using F243 [76] and R1494 [75] to generate Actinobacteria-specific PASW Statistics 18 (SPSS Inc., Chicago, USA). amplicons. The reaction mixture and PCR conditions were Epifluorescence microscopy. A cultivation-independent carried out as described by Hardoim et al. [6], except for the analysis of prokaryotic abundance based on epifluorescence concentration of Taq DNA polymerase (1.25U), number of cycles microscopy was performed in this study. For the sponge samples, (25 cycles) and extension period (1 min) in the present study. The the suspensions prepared in the abovementioned procedure were amplicons (2 mL) were used in a second PCR for DGGE analysis first centrifuged at 500 g for 2 min to remove sponge cells and using the primers F984-GC and R1378 [76] as described

PLOS ONE | www.plosone.org 12 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges previously for total bacteria, except for the number of cycles (30 software, where only those bands displaying 50% fit range or more cycles). were considered for sequencing purposes. Discriminating bands Alphaproteobacteria 16S rRNA gene fragments. The first were excised from DG-gels and re-amplified for PCR-DGGE reaction mixture of the nested PCR was prepared as described by analysis using the method of Costa et al. [80]. The resulting Gomes et al. [78], except that in the present study the primer amplicons were loaded onto DGGE with the original community concentration and Taq DNA polymerase were 0.2 mM and 1.25 DNA samples to verify their electrophoretic mobility. Excised U, respectively, and that BSA was not used in the group-specific bands that displayed the right melting behaviour when compared PCR. After initial denaturation at 94uC for 7 min, 30 cycles of with the original band in the community profiles were used as 1 min at 94uC, 1 min at 56uC and 1 min at 72uC were carried templates in another PCR amplification, in which the forward out. The reaction was finished with an extension of 10 min at primer F984 used had no GC clamp. PCR-DGGE reaction 72uC. Amplicons from the first reaction (2 mL) were used in mixtures and thermal cycles were carried out as described above. a subsequent PCR for DGGE analysis as described previously, The amplicons were then purified in Sephadex G50 columns, except for the number of cycles (25 cycles). quantified with Image LabTM Software, and subjected to PCR-DGGE profiling. DGGE assays were carried out in sequencing as above mentioned. For some excised bands, no pure a PhorU-2 gradient system (Ingeny International, Goes, The amplicon was recovered and thus a cloning procedure was Netherlands). The 16S rRNA gene amplicons generated as undertaken using the pGEM-T Vector System II Kit (Promega, explained above were applied in even concentrations onto Madison, WI) as described elsewhere [6,77,80]. Clones that polyacrylamide gels containing a 46.5 to 65% gradient of showed the same electrophoretic mobility when compared to their denaturants (100% denaturants defined as 7 M urea and 40% original band were selected for sequencing as explained above. All formamide) and a 6.2 to 9% gradient of acrylamide. Mixtures of sequences retrieved in this study were submitted to the EMBL PCR products of ten bacterial strains isolated from Sarcotragus sp. Nucleotide Sequence Database under accession numbers and Ircinia sp. (Staphylococcus sp.; Ruegeria sp.; Pseudomonas sp.; HE797930 to HE797937 for sponge CO1 sequences and Leisingera sp.; Corynebacterium sp.; Micrococcus sp.; Streptomyces sp. and HE797938 to HE797958 for PCR-DGGE bands representing Pontibacter sp.) were loaded at the edge of the gels as markers. bacterial 16S rRNA genes. Electrophoresis was performed in a 1X Tris-acetate-EDTA buffer Phylogenetic analyses. Sequences generated from sponge (pH 7.8) at 58uC and 140V for 16 h. The gels were silver stained CO1 amplification and excised bacterial bands were quality- [76] and air dried, after which digital images were obtained by inspected and edited with the Sequence Scanner software V.1 scanning. (Applied Biosystems). For bacterial DGGE bands, taxonomic Analysis of PCR-DGGE profiles. The software GelCompar assignment of sequences was performed with the seqmatch and II 5.1 (Applied Maths, Kortrijk, Belgium) was used to analyse the classifier tools of the Ribosomal Database Project II (http://rdp. PCR-DGGE profiles as recommended by Rademaker and de cme.msu.edu) at 80% confidence threshold. Closest phylogenetic Bruijn, [79]. Briefly, pairwise Pearson correlation coefficients (r) relatives were searched with the blast-n algorithm of NCBI. PCR- were calculated as a measurement of the similarity between the DGGE band sequences and their closest phylogenetic relatives community profiles. Cluster analysis was carried out with the were aligned using the SINA web aligner [81]. Aligned sequences unweighted pair group method with mathematical averages were then imported into a modified SILVA 16S rRNA gene (UPGMA) using the similarity matrix generated with the database version 102, which included all sponge-derived 16S calculated Pearson coefficients. In addition to cluster analysis, rRNA gene sequences available in early 2010 [34], using the constrained (i.e. canonical) ordination was performed with the parsimony tool as implemented in the ARB software [82]. The Canoco for Windows 4.5 software package (Microcomputer sponge database generated by Simister et al. [34] contained Power. Ithaca, NY) using a ‘‘sample6species’’ datasheet as input, phylogenetic inferences performed with long sequences in which the ‘‘species’’ data represent the presence and relative ($1200bp) using the program RAxML for all sponge-associated abundance of PCR-DGGE bands in each fingerprint, as described bacterial phyla, from which sponge-specific clusters were assigned in detail by Costa et al. [80]. This was used to infer whether sponge [34] according to the criteria described by Hentschel et al. [69]. species and seawater significantly contributed to the observed Alignments were manually checked and corrected when necessary variability in the PCR-DGGE profiles (see [6,80]). The Shannon using the ARB alignment window. The sequences generated in measure of diversity (H’), determined as H’=–g pi.logpi where pi this study were added to maximum likelihood trees inferred by represents the relative abundance of the ith category (i.e. PCR- Simister et al. [34] through the parsimony interactive tool available DGGE band) within the sample (i.e. PCR-DGGE fingerprint) was in ARB using 50% conservation filters for each of the applied to estimate the diversity of each PCR-DGGE fingerprint corresponding bacterial phyla, and their affiliation with sponge- generated in this study. The evenness (J’ = H’/Hmax) of PCR- specific phylogenetic clusters was then ascertained. From the DGGE fingerprints was calculated based on the diversity indices resulting trees, relevant sequences were selected for further obtained. Measures of richness (i.e. number of PCR-DGGE phylogenetic analyses (see below). The CO1 gene sequences from bands), diversity and evenness of PCR-DGGE fingerprints were each investigated specimen were aligned against selected sponge compared by One Way ANOVA using PASW Statistics 18 (SPSS barcoding sequences using Clustal X in the MEGA5 software [83]. Inc., Chicago, USA). Phylogenetic inferences of bacteria and sponge sequences were Identification of dominant bands in PCR-DGGE performed as described by Hardoim et al. [84]. Briefly, an profiles. Sponge-associated and seawater exclusive bands were appropriate evolutionary model for all phylogenetic trees was visually determined based on their occurrence across replicates determined using ModelGenerator version 85 [85] and found to (i.e. only bands detected in at least 3 of 4 replicates were be the general-time reversible model (GTR, [86] with a discrete sequenced). Further, their discriminating power was assessed, and gamma-distribution of among-site rate variation (C4) and a pro- only bands displaying high variance in relative abundance as portion of invariant sites (I), except for CO1 inference, in which a response to the sample classes ‘‘I. variabilis’’, ‘‘S. spinosulus’’ and invariant sites did not fit. Maximum likelihood and Bayesian ‘‘Seawater’’ were selected. Discriminating bands were revealed by MCMC analyses were conducted using RAxML (vers. 7.0.4-MPI) the species fit range function in the Canoco for windows 4.5 and MrBayes (vers. 3.2.1), respectively [87–89].

PLOS ONE | www.plosone.org 13 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

Supporting Information Acknowledgments Figure S1 Sponge species. In situ pictures of S. spinosulus (A) We acknowledge Msc. Sandra Mesquita, Dr Paulo J. Gavaia and Joa˜o C. and I. variabilis (B). Silva from the Centre of Marine Sciences (CCMar-UAlg) for their help in (TIF) epifluorescence microscopy sample preparation and analysis. We thank Dr. Christin Zachow from the Institute of Environmental Biotechnology, Graz Figure S2 Cluster analysis. Cluster analysis of PCR-DGGE University of Technology, Austria, for her help with Gelcompar analysis. fingerprints obtained for Bacteria (A), Actinobacteria (B) and We are grateful to Dr. Michael Taylor for kindly providing the ARB Alphaproteobacteria (C). S. spinosulus: Alg10/08, Alg10/09, Alg10/ database containing sponge-derived 16S rRNA gene sequences [34] used in this study to diagnose sponge-specific clusters. 10 and Alg10/11; I. variabilis: Alg10/12, Alg10/13, Alg10/14 and Alg10/15 and Seawater: SW07, SW22, SW23 and SW24. (TIF) Author Contributions Conceived and designed the experiments: JMSG JRX RC. Performed the Table S1 PCR-DGGE band richness, diversity and experiments: CCPH AISE FRP JRX. Analyzed the data: CCPH CC JRX evenness. RC. Contributed reagents/materials/analysis tools: JMSG CC JRX RC. (DOCX) Wrote the paper: CCPH JRX RC.

References 1. Taylor MW, Radax R, Steger D, Wagner M (2007) Sponge-associated 23. Wa¨tjen W, Putz A, Chovolou Y, Kampko¨tter A, Totzke F, et al. (2009) Hexa-, microorganisms: Evolution, ecology, and biotechnological potential. Microbiol hepta- and nonaprenylhydroquinones isolated from marine sponges Sarcotragus Mol Biol R 71: 295–347. muscarum and Ircinia fasciculata inhibit NF-kappa B signalling in H4IIE cells. 2. Henstchel U, Piel J, Degnan SM, Taylor MW (2012) Genomic insights into the J Pharm Pharmacol 61: 919–924. marine sponge microbiome. Nat Rev Microbiol 10: 641–654. 24. Orhan I, Sener B, Kaiser M, Brun R, Tasdemir D (2010) Inhibitory activity of 3. Vacelet J, Donadey C (1977) Electron-microscope study of association between marine sponge-derived natural products against parasitic Protozoa. Mar Drugs some sponges and bacteria. J Exp Mar Biol Ecol 30: 301–314. 8: 47–58. 4. Friedrich AB, Merkert H, Fendert T, Hacker J, Proksch P, et al. (1999) 25. De Rosa S, Mitova M, Tommonaro G (2003) Marine bacteria associated with Microbial diversity in the marine sponge Aplysina cavernicola (formerly Verongia sponge as source of cyclic peptides. Biomol Eng 20: 311–316. cavernicola) analyzed by fluorescence in situ hybridization (FISH). Mar Biol 134: 26. Mitova M, Tommonaro G, De Rosa S (2003) A novel cyclopeptide from 461–470. a bacterium associated with the marine sponge Ircinia muscarum. Z Natur- 5. Hentschel U, Usher KM, Taylor MW (2006) Marine sponges as microbial forschung C 58c: 740–745. fermenters. FEMS Microbiol Ecol 55: 167–177. 27. Fisch K, Gurgui C, Heycke N, van der Sar S, Anderson S, et al. (2009) 6. Hardoim CCP, Costa R, Araujo FV, Hajdu E, Peixoto R, et al. (2009) Diversity Polyketide assembly lines of uncultivated sponge symbionts from structure-based of bacteria in the marine sponge Aplysina fulva in Brazilian coastal waters. Appl gene targeting. Nat Chem Biol 5: 494–501. Environ Microb 75: 3331–3343. 28. Schmitt S, Tsai P, Bell J, Fromont J, Ilan M, et al. (2011) Assessing the complex 7. Thomas TRA, Kavlekar DP, LokaBharathi PA (2010) Marine drugs from sponge microbiota: core, variable and species-specific bacterial communities in sponge-microbe association - A Review. Mar Drugs 8: 1417–1468. marine sponges. ISME J 6: 564–576. 8. Webster NS, Taylor MW, Behnam F, Lu¨cker S, Rattei T, et al. (2010) Deep 29. Kvennefors ECE, Sampayo E, Ridgway T, Barnes AC, Hoegh-Guldberg O sequencing reveals exceptional diversity and modes of transmission for bacterial (2010) Bacterial communities of two ubiquitous Great Barrier Reef corals reveals sponge symbionts. Environ Microbiol 12: 2070–2082. both site-and species-specificity of common bacterial associates. PLoS ONE 5: 9. Lee OO, Wang Y, Yang JK, Lafi FF, Al-Suwailem A, et al. (2011) e10401. doi:10.1371/journal.pone.0010401. Pyrosequencing reveals highly diverse and species-specific microbial communi- 30. Fierer N, Lauber CL, Zhou N, McDonald D, Costello EK, et al. (2010) Forensic ties in sponges from the Red Sea. ISME J 5: 650–664. identification using skin bacterial communities. P Natl Acad Sci USA 107: 6477– 10. Usher K, Kuo J, Fromont J, Toze S, Sutton D (2006) Comparative morphology 6481. of five species of symbiotic and non-symbiotic coccoid Cyanobacteria. Eur J Phycol 31. Arumugam M, Raes J, Pelletier E, Le Paslier D, Yamada T, et al. (2011) 41: 179–188. Enterotypes of the human gut microbiome. Nature 473: 174–180. 11. Schmitt S, Weisz J, Lindquist N, Hentschel U (2007) Vertical transmission of 32. Bull AT, Stach JEM (2007) Marine Actinobacteria: new opportunities for natural a phylogenetically complex microbial consortium in the viviparous sponge Ircinia product search and discovery. Trends Microbiol 15: 491–499. felix. Appl Environ Microb 73: 2067–2078. 33. Webster NS, Taylor MW (2012) Marine sponges and their microbial symbionts: 12. Weisz J, Hentschel U, Lindquist N, Martens C (2007) Linking abundance and love and other relationships. Environ Microbiol 14: 335–346. diversity of sponge-associated microbial communities to metabolic differences in 34. Simister RL, Deines P, Botte´ ES, Webster NS, Taylor MW (2012) Sponge- host sponges. Mar Biol 152: 475–483. specific clusters revisited: a comprehensive phylogeny of sponge-associated 13. Mohamed NM, Colman AS, Tal Y, Hill RT (2008) Diversity and expression of microorganisms. Environ Microbiol 14: 514–524. nitrogen fixation genes in bacterial symbionts of marine sponges. Environ 35. Cook SC, Bergquist PR (2002) Family Irciniidae Gray, 1867, In: Hooper JNA, Microbiol 10: 2910–2921. van Soest RWM, editors. System Porifera: a guide to the classification of 14. Mohamed N, Cicirelli E, Kan J, Chen F, Fuqua C, et al. (2008) Diversity and sponges. Kluwer Academic/Plenum Publishers, New York, NY. 1022–1027. quorum-sensing signal production of Proteobacteria associated with marine 36. Erpenbeck D, Breeuwer J, van der Velde H, van Soest R (2002) Unravelling host sponges. Environ Microbiol 10: 75–86. and symbiont phylogenies of halichondrid sponges (Demospongiae, Porifera) 15. Mohamed N, Rao V, Hamann M, Kelly M, Hill R (2008) Monitoring bacterial using a mitochondrial marker. Mar Biol 141: 377–386. diversity of the marine sponge upon transfer into aquaculture. 37. Wo¨rheide G, Erpenbeck D (2007) DNA of sponges - progress and Appl Environ Microb 74: 4133–4143. perspectives. J Mar Biol Assoc UK 87: 1629–1633. 16. Mohamed NM, Saito K, Tal Y, Hill RT (2010) Diversity of aerobic and 38. Erpenbeck D, Duran S, Rutzler K, Paul V, Hooper JNA, et al. (2007) Towards anaerobic ammonia-oxidizing bacteria in marine sponges. ISME J 4: 38–48. a DNA taxonomy of Caribbean : a gene tree reconstructed from 17. Yang JK, Sun J, Lee OO, Wong YH, Qian PY (2011) Phylogenetic diversity and partial mitochondrial CO1 gene sequences supports previous rDNA phylogenies community structure of sponge-associated bacteria from mangroves of the and provides a new perspective on the systematics of Demospongiae. J Mar Biol Caribbean Sea. Aquat Microb Ecol 62: 231–240. Assoc UK 87: 1563–1570. 18. Erwin PM, Lo´pez-Legentil S, Gonzalez-Pech R, Turon X (2012) A specific mix 39. Cardenas P, Menegola C, Rapp HT, Diaz MC (2009) Morphological of generalists: bacterial symbionts in Mediterranean Ircinia spp. FEMS Microbiol description and DNA barcodes of shallow-water Tetractinellida (Porifera: Ecol 79: 619–637. Demospongiae) from Bocas del Toro, Panama´, with description of a new species. 19. Cichewicz R, Valeriote F, Crews P (2004) Psymberin, a potent sponge-derived Zootaxa: 1–39. cytotoxin from Psammocinia distantly related to the pederin family. Org Lett 6: 40. Po¨ppe J, Sutcliffe P, Hooper JNA, Wo¨rheide G, Erpenbeck D (2010) CO I 1951–1954. barcoding reveals new clades and radiation patterns of Indo-Pacific sponges of 20. Emura C, Higuchi R, Miyamoto T (2006) Irciniasulfonic acid B, a novel taurine the Family Irciniidae (Demospongiae: Dictyoceratida). PLoS ONE 5: e9950. conjugated fatty acid derivative from a Japanese marine sponge, Ircinia sp. doi:10.1371/journal.pone.0009950. Tetrahedron 62: 5682–5685. 41. Xavier JR, Van Soest RWM, Breeuwer JAJ, Martins AMF, Menken SBJ (2010) 21. Liu Y, Zhang S, Abreu P (2006) Heterocyclic terpenes: linear furano- and Phylogeography, genetic diversity and structure of the poecilosclerid sponge pyrroloterpenoids. Nat Prod Rep 23: 630–651. Phorbas fictitius at oceanic islands. Contrib Zool 79: 119–129. 22. Xu S, Liao X, Du B, Zhou X, Huang Q, et al. (2008) A series of new 5,6- 42. Thacker R, Starnes S (2003) Host specificity of the symbiotic cyanobacterium epoxysterols from a Chinese sponge Ircinia aruensis. Steroids 73: 568–573. Oscillatoria spongeliae in marine sponges, Dysidea spp. Mar Biol 142: 643–648.

PLOS ONE | www.plosone.org 14 December 2012 | Volume 7 | Issue 12 | e53029 Bacterial Diversity in Irciniidae Sponges

43. Montalvo NF, Hill RT (2011) Sponge-associated bacteria are strictly maintained 66. Srinivas T, Kumar PA, Sasikala C, Ramana CV (2007) Rhodovulum imhoffii sp. in two closely related but geographically distant sponge hosts. Appl Environ nov. Int J Syst Evol Micr 57: 228–232. Microb 77: 7207–7216. 67. Imhoff JF (2005) Genus XIV. Rhodovulum Hiraishi and Ueda 1994. In: Brenner 44. Santavy DL, Willenz P, Colwell RR (1990) Phenotypic study of bacteria DJ, Krieg NR, Staley NR, Garrity GM, editors. Bergeys Manual of systematic associated with the caribbean sclerosponge, Ceratoporella nicholsoni. Appl Environ Bacteriology, 2nd edn, vol. 2 (The Proteobacteria), part C (The Alphaproteobacteria, Microb 56: 1750–1762. Betaproteobacteria, Deltaproteobacteria and Epsilonproteobacteria),. New York: Springer. 45. Friedrich AB, Fischer I, Proksch P, Hacker J, Hentschel U (2001) Temporal 205–209. variation of the microbial community associated with the mediterranean sponge 68. Brusca RC, Brusca GJ (2002) Phylum Porifera: the sponges,. In Sunauer AD, Aplysina aerophoba. FEMS Microbiol Ecol 38: 105–113. editor. Invertebrates. Sinauer Associates, Inc.,Cambridge, MA. 179–208. 46. Webster NS, Hill RT (2001) The culturable microbial community of the Great 69. Hentschel U, Hopke J, Horn M, Friedrich AB, Wagner M, et al. (2002) Barrier Reef sponge Rhopaloeides odorabile is dominated by an alpha-Proteobac- Molecular evidence for a uniform microbial community in sponges from terium. Mar Biol 138: 843–851. different oceans. Appl Environ Microb 68: 4431–4440. 47. Vicente VP (1990) Response of sponges with autotrophic endosymbionts during 70. McLachlan J (1964) Some considerations of growth of marine algae in artificial the coral-bleaching episode in Puerto Rico. Coral Reefs 8: 199–202. media. Can J Microbiol 10: 769–772. 48. Weisz JB, Lindquist N, Martens CS (2008) Do associated microbial abundances 71. Meyer CP, Geller JB, Paulay G (2005) Fine scale endemism on coral reefs: impact marine pumping rates and tissue densities? Oecologia 155: Archipelagic differentiation in turbinid gastropods. Evolution 59: 113–125. 367–376. 72. Folmer O, Black M, Hoeh W, Lutz R Vrijenhoek. R. 1994. DNA primers for 49. Taylor MW, Schupp PJ, Dahllof I, Kjelleberg S, Steinberg PD (2004) Host amplification of mitochondrial cytochrome c oxidase subunit I from diverse specificity in marine sponge-associated bacteria, and potential implications for metazoan invertebrates. Mol Mar Biol Biotech 3: 294–299. marine microbial diversity. Environ Microbiol 6: 121–130. 73. Altschul SF, Madden TL, Schaffer AA, Zhang JH, Zhang Z, et al. (1997) 50. Taylor MW, Schupp PJ, de Nys R, Kjelleberg S, Steinberg PD (2005) Gapped BLAST and PSI-BLAST: a new generation of protein database search Biogeography of bacteria associated with the marine sponge Cymbastela concentrica. programs. Nucleic Acids Res 25: 3389–3402. Environ Microbiol 7: 419–433. 74. Garson MJ, Flowers AE, Webb RI, Charan RD, McCaffrey EJ (1998) A 51. Fan L, Reynolds D, Liu M, Stark M, Kjelleberg S, et al. (2012) Functional sponge/dinoflagellate association in the haplosclerid sponge Haliclona sp.: cellular equivalence and evolutionary convergence in complex communities of microbial origin of cytotoxic alkaloids by Percoll density gradient fractionation. Cell Tissue sponge symbionts. P Natl Acad Sci USA 109: E1878–E1887. Res 293: 365–373. 52. Stackebrandt E, Rainey FA, Ward-Rainey NL (1997) Proposal for a new 75. Weisburg WG, Barns SM, Pelletier DA, Lane DJ (1991) 16S ribosomal DNA hierarchic classification system, Actinobacteria classis nov. Int J Syst Bacteriol 47: amplification for phylogenetic study. J Bacteriol 173: 697–703. 479–491. 76. Heuer H, Krsek M, Baker P, Smalla K, Wellington EMH (1997) Analysis of 53. Lee OO, Chui PY, Wong YH, Pawlik JR, Qian PY (2009) Evidence for vertical actinomycete communities by specific amplification of genes encoding 16S transmission of bacterial symbionts from adult to embryo in the Caribbean rRNA and gel-electrophoretic separation in denaturing gradients. Appl Environ sponge Svenzea zeai. Appl Environ Microb 75: 6147–6156. Microb 63: 3233–3241. 54. Schmitt S, Angermeier H, Schiller R, Lindquist N, Hentschel U (2008) 77. Costa R, Go¨tz M, Mrotzek N, Lottmann J, Berg G, et al. (2006) Effects of site Molecular microbial diversity survey of sponge reproductive stages and and plant species on rhizosphere community structure as revealed by molecular mechanistic insights into vertical transmission of microbial symbionts. Appl analysis of microbial guilds. FEMS Microbiol Ecol 56: 236–249. Environ Microb 74: 7694–7708. 78. Gomes NCM, Heuer H, Schonfeld J, Costa R, Mendonc¸a -Hagler L, et al. 55. Clark DA, Norris PR (1996) Acidimicrobium ferrooxidans gen nov, sp nov: Mixed- (2001) Bacterial diversity of the rhizosphere of maize (Zea mays) grown in tropical culture ferrous iron oxidation with Sulfobacillus species. Microbiol UK 142: 785– soil studied by temperature gradient gel electrophoresis. Plant Soil 232: 167–180. 790. 79. Rademaker J, de Bruijn FJ (2004) Computer-assisted analysis of molecular 56. Clum A, Nolan M, Lang E, Del Rio TG, Tice H, et al. (2009) Complete genome fingerprint profiles and database construction, In: GA Kowalchuk, FJ de Bruijn, sequence of Acidimicrobium ferrooxidans type strain (ICP(T)). Stand Genomic Sci 1: IM Head, AD Akkermans, JD van Elsas editors. Molecular Microbial Ecology 38–45. Manual. Kluwer Academic Publishers, Dordrecht, The Netherlands. 1397– 57. Davis-Belmar CS, Norris PR (2009) Ferrous iron and pyrite oxidation by 1446. "Acidithiomicrobium" species. Adv Mater Res 71–73: 271–274. 80. Costa R, Salles JF, Berg G, Smalla K (2006) Cultivation-independent analysis of 58. Johnson DB, Bacelar-Nicolau P, Okibe N, Thomas A, Hallberg KB (2009) Pseudomonas species in soil and in the rhizosphere of field-grown Verticillium dahliae Ferrimicrobium acidiphilum gen. nov., sp. nov. and Ferrithrix thermotolerans gen. nov., host plants. Environ Microbiol 8: 2136–2149. sp. nov.: heterotrophic, iron-oxidizing, extremely acidophilic Actinobacteria. 81. Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig WG, et al. (2007) SILVA: Int J Syst Evol Micr 59: 1082–1089. a comprehensive online resource for quality checked and aligned ribosomal 59. Imhoff JF (2005) Order I. Chromatiales ord. nov. In: Brenner DJ, Krieg NR, RNA sequence data compatible with ARB. Nucleic Acids Res 35: 7188–7196. Staley NR, Garrity GM, editors. Bergeys Manual of Systematic Bacteriology, 2nd 82. Ludwig W, Strunk O, Westram R, Richter L, Meier H, et al. (2004) ARB: edn, vol. 2 (The Proteobacteria), part B (The Gammaproteobacteria). New York: a software environment for sequence data. Nucleic Acids Res 32: 1363–1371. Springer. 1–3. 83. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, et al. (2011) MEGA5: 60. Proctor LM (1997) Nitrogen-fixing, photosynthetic, anaerobic bacteria associ- molecular evolutionary genetics analysis using maximum likelihood, evolution- ated with pelagic copepods. Aquat Microb Ecol 12: 105–113. ary distance, and maximum parsimony methods. Mol Biol Evol 28: 2731–2739. 61. Schmitt S, Deines P, Behnam F, Wagner M, Taylor MW (2011) Chloroflexi 84. Hardoim CCP, Cox CJ, Peixoto RS, Rosado AS, Costa R, et al. (2012) Diversity bacteria are more diverse, abundant, and similar in high than in low microbial of the candidate phylum Poribacteria in the marine sponge Aplysina fulva. Braz J abundance sponges. FEMS Microbiol Ecol 78: 497–510. Microbiol. In press. 62. Bruck WM, Bruck TB, Self WT, Reed JK, Nitecki SS, et al. (2010) Comparison 85. Keane TM, Creevey CJ, Pentony MM, Naughton TJ, McInerney JO (2006) of the anaerobic microbiota of deep-water Geodia spp. and sandy sediments in the Assessment of methods for amino acid matrix selection and their use on Straits of Florida. ISME J 4: 686–699. empirical data shows that ad hoc assumptions for choice of matrix are not 63. Siegl A, Hentschel U (2010) PKS and NRPS gene clusters from microbial justified. BMC Evol Biol 6: 29. symbiont cells of marine sponges by whole genome amplification. Environ 86. Rodriguez F, Oliver JL, Marin A, Medina JR (1990) The general stochastic- Microbiol Rep 2: 507–513. model of nucleotide substitution. J Theor Biol 142: 485–501. 64. Garrity (2006) Order III. Rhodobacterales ord. nov. In: Brenner DJ, Krieg NR, 87. Huelsenbeck JP, Ronquist F (2001) MrBayes: Bayesian inference of phylogenetic Staley NR, Garrity GM, editors. Bergeys Manual of systematic Bacteriology, 2nd trees. Bioinformatics 17: 754–755. edn, vol. 2 (The Proteobacteria), part C (The Alphaproteobacteria, Betaproteobacteria, 88. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference Deltaproteobacteria and Epsilonproteobacteria). New York: Springer. 161–167. under mixed models. Bioinformatics 19: 1572–1574. 65. Hiraishi A, Ueda Y (1995) Isolation and characterization of Rhodovulum strictum 89. Stamatakis A (2006) RAxML-VI-HPC: Maximum likelihood-based phylogenetic sp. nov. and some other purple nonsulfur bacteria from colored blooms in tidal analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688– and seawater pools. Int J Syst Bacteriol 45: 319–326. 2690.

PLOS ONE | www.plosone.org 15 December 2012 | Volume 7 | Issue 12 | e53029